The impact of turbulence on star formation and supermassive black hole growth at high redshift

在推动星系形成和演化中发挥关键作用的物理过程，如恒星形成和恒星反馈，以及超大质量黑洞形成、吸积和反馈，都发生在极小的亚星系尺度上。因此，数值流体动力学模拟既要捕捉星系的环境，又要有意义地解析它们的内部结构，以确定恒星形成的位置，这是一个巨大的挑战。特别是，可压缩湍流在形成星系特性方面所起的作用受到的关注相对较少，尽管人们早就知道（Larson, 1981）恒星形成的摇篮，即分子云，是超音速湍流。因此，该项目的目的是试图弥合这一差距，并确定驱动星系湍流的相关物理过程，同时量化这种湍流如何塑造星系的星际介质并影响恒星/黑洞的形成。更具体地说，这包括运行/分析单个星系在其明确的宇宙学背景下的高分辨率变焦（最小10倍）模拟，以解决环境（大规模气体流动，飞掠相遇，合并）在驱动/触发湍流中所起作用的复杂问题，并明确使用部分解决的湍流级联，在模拟中实时测量。确定恒星（黑洞）形成的位置和恒星形成过程的效率。然后将这些模拟与它们的“双胞胎”进行比较，即在其他方面都相同的模拟，除了恒星形成是由当前的“标准”子网格模型决定的，即一个简单的密度阈值和一个固定的，普遍的效率，根据观察到的Kennicutt- schmidt定律进行经验校准（Kennicutt, 1998）。最终，湍流恒星形成模型的成功与否将取决于其恢复kennicut - schmidt定律（与标准模型相反，该定律不是内置的）的能力，同时产生具有与观测值一致的暗物质与恒星质量比的星系（Moster et al 2012, Behroozi et al 2013）。

Physical processes that play a key role in driving galaxy formation and evolution, such as star formation and stellar feedback, and super massive black hole formation, accretion and feedback, take place on extremely small, sub-galactic scales. For this reason, it is a tremendous challenge for numerical hydrodynamics simulations to capture both the environment of galaxies and meaningfully resolve their internal structure to identify the sites where stars form. In particular, the role played by compressible turbulence in shaping galaxy properties has received relatively little attention, even though it has long been known (Larson, 1981) that the cradles of star formation, i.e. molecular clouds, are supersonically turbulent. The aim of the project is therefore to try to bridge this gap and identify the relevant physical processes which drive turbulence in galaxies whilst also quantifying how this turbulence shapes the interstellar medium of galaxies and influences star/black hole formation. More specifically, this involves running/analysing high-resolution zoom (min 10pc) simulations of individual galaxies in their explicit cosmological context, to address the complex issue of the role played by the environment (large scale gas flows, fly-by encounters, mergers) in driving/triggering turbulence, and explicitly use the partially resolved turbulent cascade of the flow, measured on-the-fly in the simulations, to determine both the loci of star (black hole) formation and the efficiency of the star formation process. These simulations will then be compared to their "twins", i.e. simulations that are identical in every other aspect, except that star formation is determined by the current "standard" sub-grid modelling, i.e. a simple density threshold and a fixed, universal efficiency, empirically calibrated on the observed Kennicutt-Schmidt law (Kennicutt, 1998). Ultimately, the success of the turbulent star formation model will be judged by its capacity to recover the Kennicutt-Schmidt law (which, contrary to the standard model, is not built-in) while at the same time producing galaxies with dark matter to stellar mass ratios in agreement with observed values (Moster et al 2012, Behroozi et al 2013).